341 commits

Author	SHA1	Message	Date
alemi	4b4d52ef6f	feat: fetch related things ... fetching an object will get its author fetching an activity will get its object and author	2024-04-23 17:56:46 +02:00
alemi	597dee934d	fix: try hiding locked activities for public objs ... idk if this works but basically when there's a public object from a private activity, the query joins the activity anyway because it uses the object relation, try using the addressing relation and see what comes out	2024-04-23 17:21:32 +02:00
alemi	8a6c05f4fb	fix: relay follow/accept are public	2024-04-23 16:54:03 +02:00
alemi	2a9cf5bcf5	feat: command to also accept relay follow reqs ... this has quite awful UX because you can't really know if relay is following back without checking logs, but whatevvvv works for now	2024-04-23 16:26:19 +02:00
alemi	dfbadf5d11	fix: oops i meant server inbox GET, now it is	2024-04-23 06:29:52 +02:00
alemi	2bf4469270	feat: server timeline is object-based	2024-04-23 06:21:02 +02:00
alemi	70920985cf	fix: use hs2019 instead of rsa-sha256 ... im not sure why but apparently there's some bug somewhere? maybe some instances are picky and want the new thing? should still be fine using sha256 as signing tho	2024-04-23 05:46:48 +02:00
alemi	1731bcb1c6	fix: remove again trailing slash	2024-04-23 05:31:19 +02:00
alemi	2cb6cfd2f1	fix: well-known nodeinfo base url	2024-04-23 05:29:26 +02:00
alemi	8b88d24250	fix: show signature when invalid	2024-04-23 05:19:24 +02:00
alemi	e316e1a840	fix: also change main key name for actor	2024-04-23 05:07:04 +02:00
alemi	4f8132a642	fix: dispatcher check doesn't include trailing	2024-04-23 05:01:59 +02:00
alemi	8150b9082d	fix: proper addressing, ctx.base() with trailing / ... maybe aode-relay (and rust IRI parser) goes mad if there's no trailing slash? idk, i'm speculating	2024-04-23 04:56:29 +02:00
alemi	63d78e7ff5	fix: delete with filters instead of activemodel	2024-04-23 04:24:27 +02:00
alemi	43d5571b27	feat: better error mappings to http	2024-04-23 04:07:32 +02:00
alemi	377ed9665d	fix: set published before converting to model ... user may not have set it and we overwrite it anyway. id instead is necessary, so if it's not given we should return 500	2024-04-23 03:59:27 +02:00
alemi	867740b2e6	fix: actually hook up all outbox actions	2024-04-23 03:45:50 +02:00
alemi	852b45f8dd	fix: also fix context for fetched objects	2024-04-23 03:04:54 +02:00
alemi	46de9aebd6	feat: implemented update, delete, announce outbox	2024-04-22 22:52:19 +02:00
alemi	8dece0f004	feat: implemented authed inbox	2024-04-22 22:52:05 +02:00
alemi	3c6b2ae3e6	fix: don't embed remote objects when dispatching ... weak check but should still save some bytes	2024-04-22 22:42:00 +02:00
alemi	b62b461305	fix: dispatcher fetch app key for app deliveries	2024-04-22 04:23:11 +02:00
alemi	1d8b69dda7	fix: try setting digest always ... aode relay complains that digest is missing on fetches? idk, let's try putting an empty digest, will aode work? will mastodon/akkoma still work? will this fix some *keys too???	2024-04-22 04:11:59 +02:00
alemi	4d0fb9b684	fix: pass actor, inbox gets resolved automatically	2024-04-22 04:03:47 +02:00
alemi	32707bda0c	feat: added command to follow relays	2024-04-22 03:56:07 +02:00
alemi	87a5a77e19	chore: imports	2024-04-22 02:52:18 +02:00
alemi	f2f7b51fdc	feat: order by publish date, not discovery ... also centralized ordering inside select_* queries	2024-04-22 02:49:04 +02:00
alemi	b6f10e9cc1	fix: webfinger for application returns itself	2024-04-22 01:51:37 +02:00
alemi	dfec715594	chore: comments, clearer code	2024-04-22 01:01:42 +02:00
alemi	e9a19b3cb4	feat: fetch threads recursively up to 16	2024-04-22 00:43:17 +02:00
alemi	d575ef0bef	fix: add endpoints and preferred username to app	2024-04-22 00:37:21 +02:00
alemi	e8dd15f8bd	fix: also context gives full objects ... would clear local cached image oof	2024-04-22 00:16:40 +02:00
alemi	96524dbd6f	fix: cheap fix for mastodon contexts	2024-04-22 00:01:09 +02:00
alemi	cc287d3aa4	feat: show attachments in inbox, outbox and /object ... attachments are lazy loaded, so it may be efficient if not all posts have media, but it should probably be eager loaded anyway eventually	2024-04-21 23:22:55 +02:00
alemi	c595f5f5e3	feat: process and store remote attachment urls	2024-04-21 22:58:33 +02:00
alemi	cd9d3f9db4	fix: use preferred_username before using url	2024-04-21 21:15:58 +02:00
alemi	624492bcfb	feat: introduced arbitrary fetch route for users	2024-04-21 19:15:47 +02:00
alemi	48bba8af40	feat: root url redirects browsers to /web	2024-04-21 19:01:51 +02:00
alemi	4e5e8f8532	fix: use WrappedObject type for context query	2024-04-21 17:11:50 +02:00
alemi	645e14cc32	fix: ops wrong relation direction	2024-04-21 16:39:56 +02:00
alemi	a083edef6e	fix: sqlite won't update columns anyway ... since this migration breaks all sqlite dbs, i changed the original migration so that future ones won't panic when reaching here. note that, if you are on sqlite, just `sqlite3 <your_db> .dump > backup.sql` and then, after rebuilding and re-migrating db, `cat backup.sql | sqlite3 <your_db>`	2024-04-21 16:30:29 +02:00
alemi	46a2e53da0	fix: use wrapped object since we ask for more cols	2024-04-21 16:14:47 +02:00
alemi	5d501a3018	fix: renamed methods	2024-04-21 15:42:15 +02:00
alemi	379b4daa84	fix: don't create mock activities ... basically remade addressing table so that activity too can be null	2024-04-21 15:41:29 +02:00
alemi	fa74c3dd0a	chore: imports	2024-04-20 04:34:47 +02:00
alemi	6c3aead68b	feat: handle announces	2024-04-20 04:33:23 +02:00
alemi	f3dcbffeca	feat: parse object comments count from replies	2024-04-20 04:26:31 +02:00
alemi	1ca97668a1	chore: new helper, renamed base->domain	2024-04-20 04:26:16 +02:00
alemi	4d4cbe0ef8	feat: create mock activities for fetched objects	2024-04-20 04:24:48 +02:00
alemi	2d10f29b56	fix: restore federation, remove embedded follow___ ... it was so good, why make **just these** only links?????	2024-04-19 18:31:20 +02:00